> We are still in the waterwheel phase of AI, bolting chatbots onto workflows designed for humans. We need to imagine what knowledge work could look like when human organizations are reinforced with steel, when busywork is delegated to minds that never sleep.
【📣論文採択ニュース】
メルカリAIチームによる論文 "Towards Better Search with Domain-Aware Text Embeddings for C2C Marketplaces" が、国際会議「#AAAI2026」のthe New Frontiers in Information Retrieval ワークショップセッションに採択されました!
frontier-ir-workshop.github.io
Not a big surprise that tuned embedding models work but nevertheless.
Mercari fine-tuned tembeddings on purchase data. Significant revenue lift in A/B tests.
Generic off the shelf embeddings are leaving money on the table for domain-specific search. arxiv.org/abs/2512.21021
Thanks again all for a wonderful #EMNLP2025! For future reference, you can now find all accepted papers, award winners, and outstanding reviewers, area chairs, and senior area chairs on the conference website: 2025.emnlp.org/program/awards/
We keep scaling model parameters by increasing width and stacking more layers, but what if the truly missing axes for continual learning are compression and stacking the learning process?
Excited to share the full version of Nested Learning, a new paradigm for continual learning and machine learning in general.
Paper: nestedlearning.net/paper
Today we’re excited to unveil a new generation of Segment Anything Models:
1️⃣ SAM 3 enables detecting, segmenting and tracking of objects across images and videos, now with short text phrases and exemplar prompts.
🔗 Learn more about SAM 3: go.meta.me/591040
2️⃣ SAM 3D brings the model collection into the 3rd dimension to enable precise reconstruction of 3D objects and people from a single 2D image.
🔗 Learn more about SAM 3D: go.meta.me/305985
These models offer innovative capabilities and unique tools for developers and researchers to create, experiment and uplevel media workflows.
Meet SAM 3, a unified model that enables detection, segmentation, and tracking of objects across images and videos. SAM 3 introduces some of our most highly requested features like text and exemplar prompts to segment all objects of a target category.
Learnings from SAM 3 will help power new features in Instagram Edits and Vibes, bringing advanced segmentation capabilities directly to creators.
🔗 Learn more: go.meta.me/591040
昨日11月13日、無事にメルカリ主催のテックイベント「mercari Gears 2025」を開催することができました。ご来場いただいた多くの皆さまに、心より感謝申し上げます。
私たちにとっても7年ぶりのリアル開催ということで、少なからず緊張もありましたが、参加された皆さまがとても温かく、積極的にコミュニケーションを取ってくださったおかげで、素晴らしいカンファレンスとなりました。
AIの進化が目覚ましく、正解が定まらない時代の中で、各領域のリーダーたちが「今後のビジョン」を整理し、共有し、それに対して参加者の皆さまから直接フィードバックをいただけたことは、非常に貴重な経験でした。改めて、リアルイベントの持つ力と意義を強く感じる機会となりました。
同時に、メルカリが「AI Native Company」へと進化していくためには、アーキテクチャやプラットフォームの整備など、まだまだ並行して進めていくべき課題が多いことも再認識しました。
この大きな変化の時期だからこそ、業界を越えて語り合い、互いに学び合いながら、より高いスタンダードを共に築いていける場をこれからも増やしていきたいと感じています。
当日の発表内容は、後日動画として公開予定です。ぜひご覧いただければ幸いです。
こちらは当日のKeynoteのプレゼンテーションのリンクとなります。
speakerdeck.com/mercari/mercar…
----
Hi! Yesterday, on November 13th, we successfully held mercari Gears 2025, Mercari’s tech event. I want to express my gratitude to everyone who joined us.
It was our first in-person tech conference in seven years, and while we were admittedly a bit nervous, the warm energy and active engagement from all participants made it a truly inspiring and memorable event.
In this era of rapid AI evolution—where there are no absolute answers—our leaders across different domains shared their visions for the future. Receiving direct feedback from attendees was incredibly valuable, reminding us once again of the power and importance of real, face-to-face interactions.
At the same time, we were reminded that becoming an AI Native Company requires continued effort to evolve our architecture, strengthen our platforms, and address many challenges in parallel.
Especially in a period of such transformation, I believe it’s crucial for us to keep creating spaces where people across industries can come together, exchange ideas, and raise the collective standard of our ecosystem.
The session recordings will be released soon — I hope you’ll take a moment to watch them once they’re available. Thanks!
Here is the link to the keynote.
speakerdeck.com/mercari/mercar…
Meta just released new multilingual MetaCLIP2 models supporting 300+ models 🌎
you can use these models to do image ⇆ text search in databases in any language! 🔍
on the next one is a notebook on how to do so 🫡